Atlas / Skills / jeremylongshore / Instantly Local Dev Loop

Instantly Local Dev LoopSAFE

skills/jeremylongshore/instantly-local-dev-loop

Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.

Verdict
SAFE
Grade
B
Trust score
89 /100
Version
1.13.0
Hosts
1 documented
License
MIT
Stars
2,824
01

Overview

Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.

Read from source at commit 4f83675ca38aOBSERVED · 2026-10-09
02

Host compatibility

What the documentation claims. We have not run a compatibility test.

HostStatusNotes
claude-codementioned
03

What it tells the agent

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: instantly-local-dev-loop
description: >-
  Build an offline-first Instantly development loop with fixtures, fakes, and optional read-only staging probes. Use when developers need fast repeatable tests without mutating a live workspace. Trigger with "mock the Instantly API", "build Instantly fixtures", or "test Instantly locally".
argument-hint: "[repository-path] [offline|staging]"
allowed-tools: Read, Glob, Grep, WebFetch, Write, Edit
version: 1.13.0
author: Jeremy Longshore <[email protected]>
license: MIT
tags:
- saas
- instantly
- local-dev-loop
model: inherit
effort: high
compatibility: Designed for Claude Code; live verification requires network access and an approved Instantly workspace and API v2 key
---
# Instantly Local Contract Loop

## Overview

Develop safely without relying on an undocumented vendor mock endpoint or sending real outreach. Record assumptions, evidence, approval state, and rollback ownership so another operator can reproduce the result.

## Prerequisites

- The target repository, Instantly workspace, environment, and accountable owner
- Current security, privacy, compliance, capacity, and change-control requirements
- An approved API v2 key only when a bounded live verification is necessary

## Tool Discipline

Use `Read`, `Glob`, and `Grep` to inspect code, configuration, and evidence. Use `WebFetch` only for current first-party Instantly documentation and package metadata. Use `Write` or `Edit` only when implementation was requested and exact target files are known; never write credentials, lead data, email content, or unrestricted environment output.

## Current Contract

- The documentation exposes interactive examples but does not establish a durable public mock API contract.
- Use local fakes and captured synthetic schemas for deterministic tests.
- Live probes require a staging workspace, read-only scopes, and bounded requests.

## Authentication

Use an API v2 key as `Authorization: Bearer <key>` against `https://api.instantly.ai/api/v2`. Grant only the endpoint-specific scopes needed, inject the key from an approved server-side secret manager, and never print, persist, commit, or place it in a URL. Treat key creation, rotation, revocation, member changes, workspace delegation, and production access as owner-approved actions.

## Instructions

1. Inventory routes and generate or pin types from the official OpenAPI surface.
2. Create local fake responses for success, 401, 403, 429, pagination, and malformed data.
3. Use synthetic leads and messages; never copy production payloads into fixtures.
4. Run unit and contract tests without network access.
5. Optionally run one owner-approved read-only staging probe.
6. Update fixtures only after reviewed schema evidence and record the source date.

## Approval Boundaries

Do not create, rotate, reveal, or revoke keys; invite or remove members; delegate across workspaces; connect sending accounts; create or activate campaigns; import or delete leads; change suppression or retention; register, patch, resume, or delete webhooks; alter plans or paid capacity; transmit diagnostics; or perform another production mutation without explicit approval from the accountable owner. Keep diagnosis read-only unless implementation was requested.

## Output

Return the workspace-safe scope, files and contracts inspected, exact API v2 routes and required scopes, evidence collected, validation result, sensitive fields redacted, remaining risk, accountable owner, approval state, and rollback or next action.

## Error Handling

| Condition | Response |
|---|---|
| `401` | Stop and verify that the bearer key exists, is current, and was not revoked. |
| `403` | Stop and compare the operation with its exact required scope; do not broaden to `all:all` by default. |
| `429` | Coordinate the workspace-wide budget, honor endpoint overrides, and bound retries. |
| Schema or tenant mismatch | Fail closed, preserve redacted evidence, and do not retry a mutation. |

## Examples

Use a compact handoff that makes scope, mutation authority, and evidence reviewable.

Input:

```text
mode=offline; fixture=campaign-list; network=false
```

Expected handoff:

```text
unit=pass; contract=pass; staging=skipped
```

## Resources

- [Skill-specific official documentation](references/official-docs.md)
- [Instantly API v2 documentation](https://developer.instantly.ai/)
- [Instantly API v2 OpenAPI document](https://api.instantly.ai/openapi/api_v2.json)
04

Trust audit

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

LayerWhat it checksResult
L0Provenance & inventoryPASS
L1Static analysis of the codePASS
L2Instruction surface (what it tells the agent)PASS
L3Class-specific surfacePASS
L4Behavioural (sandbox)SKIPPED

What the source does

Filesystem
none-observed
Network
none-observed
Shell
none-observed
Dependencies
pinned
Secrets in source
none-found

Findings (1)

LOWPrompt injection · prompt.transfer_instruction · CWE-94, CWE-1427
references/official-docs.md:31
- Webhook events may include lead identifiers and full email/reply content. The public event guide documents payload fields and event types but not a signature secret or fixed retry schedule; do not i
Why it matters. an instruction to move sensitive data to an outside destination
Fix. remove; a skill never needs the user's secrets off the machine

Gates applied: no_behavioural_pass.

Audited 2026-10-09 · audit v0.4.1 · source sha 4f83675ca38afull audit observations/trust-audit/skill/jeremylongshore__instantly-local-dev-loop.json · Report an issue / request a re-scan
05

Audit history

Every audit this skill has had.

DateSourceVerdictGradeScoreChange
2026-10-094f83675ca38aSAFEB89first audit
06

Questions

What does the Instantly Local Dev Loop skill do?

Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.

Is Instantly Local Dev Loop safe to install?

The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean skill reads B.

What can Instantly Local Dev Loop access on my machine?

The audit observed no filesystem, network or shell use at all in its source.

Which assistants does Instantly Local Dev Loop work with?

Its documentation mentions claude-code. That is what the text claims, not a compatibility test we ran.

How current is this page?

The grade is for one exact copy of the source (4f83675ca38a), read on 2026-10-09. The repository is watched, and a new audit runs when it changes — this is the first audit.

Advertisement